error-injection: Separate error-injection from kprobe

Since error-injection framework is not limited to be used
by kprobes, nor bpf. Other kernel subsystems can use it
freely for checking safeness of error-injection, e.g.
livepatch, ftrace etc.
So this separate error-injection framework from kprobes.

Some differences has been made:

- "kprobe" word is removed from any APIs/structures.
- BPF_ALLOW_ERROR_INJECTION() is renamed to
  ALLOW_ERROR_INJECTION() since it is not limited for BPF too.
- CONFIG_FUNCTION_ERROR_INJECTION is the config item of this
  feature. It is automatically enabled if the arch supports
  error injection feature for kprobe or ftrace etc.

Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
Reviewed-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

3 files changed, 218 insertions, 0 deletions

diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
index 9d5b78aad4c5..2a33efdd1fea 100644
--- a/lib/Kconfig.debug
+++ b/lib/Kconfig.debug
@@ -1500,6 +1500,10 @@ config FAULT_INJECTION
 	  Provide fault-injection framework.
 	  For more details, see Documentation/fault-injection/.
 
+config FUNCTION_ERROR_INJECTION
+	def_bool y
+	depends on HAVE_FUNCTION_ERROR_INJECTION && KPROBES
+
 config FAILSLAB
 	bool "Fault-injection capability for kmalloc"
 	depends on FAULT_INJECTION
diff --git a/lib/Makefile b/lib/Makefile
index a6c8529dd9b2..75ec13778cd8 100644
--- a/lib/Makefile
+++ b/lib/Makefile
@@ -149,6 +149,7 @@ obj-$(CONFIG_NETDEV_NOTIFIER_ERROR_INJECT) += netdev-notifier-error-inject.o
 obj-$(CONFIG_MEMORY_NOTIFIER_ERROR_INJECT) += memory-notifier-error-inject.o
 obj-$(CONFIG_OF_RECONFIG_NOTIFIER_ERROR_INJECT) += \
 	of-reconfig-notifier-error-inject.o
+obj-$(CONFIG_FUNCTION_ERROR_INJECTION) += error-inject.o
 
 lib-$(CONFIG_GENERIC_BUG) += bug.o
 
diff --git a/lib/error-inject.c b/lib/error-inject.c
new file mode 100644
index 000000000000..bccadcf3c981
--- /dev/null
+++ b/lib/error-inject.c
@@ -0,0 +1,213 @@
+// SPDX-License-Identifier: GPL-2.0
+// error-inject.c: Function-level error injection table
+#include <linux/error-injection.h>
+#include <linux/debugfs.h>
+#include <linux/kallsyms.h>
+#include <linux/kprobes.h>
+#include <linux/module.h>
+#include <linux/mutex.h>
+#include <linux/list.h>
+#include <linux/slab.h>
+
+/* Whitelist of symbols that can be overridden for error injection. */
+static LIST_HEAD(error_injection_list);
+static DEFINE_MUTEX(ei_mutex);
+struct ei_entry {
+	struct list_head list;
+	unsigned long start_addr;
+	unsigned long end_addr;
+	void *priv;
+};
+
+bool within_error_injection_list(unsigned long addr)
+{
+	struct ei_entry *ent;
+	bool ret = false;
+
+	mutex_lock(&ei_mutex);
+	list_for_each_entry(ent, &error_injection_list, list) {
+		if (addr >= ent->start_addr && addr < ent->end_addr) {
+			ret = true;
+			break;
+		}
+	}
+	mutex_unlock(&ei_mutex);
+	return ret;
+}
+
+/*
+ * Lookup and populate the error_injection_list.
+ *
+ * For safety reasons we only allow certain functions to be overridden with
+ * bpf_error_injection, so we need to populate the list of the symbols that have
+ * been marked as safe for overriding.
+ */
+static void populate_error_injection_list(unsigned long *start,
+					  unsigned long *end, void *priv)
+{
+	unsigned long *iter;
+	struct ei_entry *ent;
+	unsigned long entry, offset = 0, size = 0;
+
+	mutex_lock(&ei_mutex);
+	for (iter = start; iter < end; iter++) {
+		entry = arch_deref_entry_point((void *)*iter);
+
+		if (!kernel_text_address(entry) ||
+		    !kallsyms_lookup_size_offset(entry, &size, &offset)) {
+			pr_err("Failed to find error inject entry at %p\n",
+				(void *)entry);
+			continue;
+		}
+
+		ent = kmalloc(sizeof(*ent), GFP_KERNEL);
+		if (!ent)
+			break;
+		ent->start_addr = entry;
+		ent->end_addr = entry + size;
+		ent->priv = priv;
+		INIT_LIST_HEAD(&ent->list);
+		list_add_tail(&ent->list, &error_injection_list);
+	}
+	mutex_unlock(&ei_mutex);
+}
+
+/* Markers of the _error_inject_whitelist section */
+extern unsigned long __start_error_injection_whitelist[];
+extern unsigned long __stop_error_injection_whitelist[];
+
+static void __init populate_kernel_ei_list(void)
+{
+	populate_error_injection_list(__start_error_injection_whitelist,
+				      __stop_error_injection_whitelist,
+				      NULL);
+}
+
+#ifdef CONFIG_MODULES
+static void module_load_ei_list(struct module *mod)
+{
+	if (!mod->num_ei_funcs)
+		return;
+
+	populate_error_injection_list(mod->ei_funcs,
+				      mod->ei_funcs + mod->num_ei_funcs, mod);
+}
+
+static void module_unload_ei_list(struct module *mod)
+{
+	struct ei_entry *ent, *n;
+
+	if (!mod->num_ei_funcs)
+		return;
+
+	mutex_lock(&ei_mutex);
+	list_for_each_entry_safe(ent, n, &error_injection_list, list) {
+		if (ent->priv == mod) {
+			list_del_init(&ent->list);
+			kfree(ent);
+		}
+	}
+	mutex_unlock(&ei_mutex);
+}
+
+/* Module notifier call back, checking error injection table on the module */
+static int ei_module_callback(struct notifier_block *nb,
+			      unsigned long val, void *data)
+{
+	struct module *mod = data;
+
+	if (val == MODULE_STATE_COMING)
+		module_load_ei_list(mod);
+	else if (val == MODULE_STATE_GOING)
+		module_unload_ei_list(mod);
+
+	return NOTIFY_DONE;
+}
+
+static struct notifier_block ei_module_nb = {
+	.notifier_call = ei_module_callback,
+	.priority = 0
+};
+
+static __init int module_ei_init(void)
+{
+	return register_module_notifier(&ei_module_nb);
+}
+#else /* !CONFIG_MODULES */
+#define module_ei_init()	(0)
+#endif
+
+/*
+ * error_injection/whitelist -- shows which functions can be overridden for
+ * error injection.
+ */
+static void *ei_seq_start(struct seq_file *m, loff_t *pos)
+{
+	mutex_lock(&ei_mutex);
+	return seq_list_start(&error_injection_list, *pos);
+}
+
+static void ei_seq_stop(struct seq_file *m, void *v)
+{
+	mutex_unlock(&ei_mutex);
+}
+
+static void *ei_seq_next(struct seq_file *m, void *v, loff_t *pos)
+{
+	return seq_list_next(v, &error_injection_list, pos);
+}
+
+static int ei_seq_show(struct seq_file *m, void *v)
+{
+	struct ei_entry *ent = list_entry(v, struct ei_entry, list);
+
+	seq_printf(m, "%pf\n", (void *)ent->start_addr);
+	return 0;
+}
+
+static const struct seq_operations ei_seq_ops = {
+	.start = ei_seq_start,
+	.next  = ei_seq_next,
+	.stop  = ei_seq_stop,
+	.show  = ei_seq_show,
+};
+
+static int ei_open(struct inode *inode, struct file *filp)
+{
+	return seq_open(filp, &ei_seq_ops);
+}
+
+static const struct file_operations debugfs_ei_ops = {
+	.open           = ei_open,
+	.read           = seq_read,
+	.llseek         = seq_lseek,
+	.release        = seq_release,
+};
+
+static int __init ei_debugfs_init(void)
+{
+	struct dentry *dir, *file;
+
+	dir = debugfs_create_dir("error_injection", NULL);
+	if (!dir)
+		return -ENOMEM;
+
+	file = debugfs_create_file("list", 0444, dir, NULL, &debugfs_ei_ops);
+	if (!file) {
+		debugfs_remove(dir);
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+static int __init init_error_injection(void)
+{
+	populate_kernel_ei_list();
+
+	if (!module_ei_init())
+		ei_debugfs_init();
+
+	return 0;
+}
+late_initcall(init_error_injection);